Fluidfont,

Tom provides a good example. Most measures of dispersion: variance, standard deviation, etc. use as part of their formula a sum of squares, or Root Mean Square. This is to help "level out" extreme values and makes it a better description of variability (you should have a good understanding of the central limit theorem) than range which only uses two values.

NOTE: Kal Kapur was an instructor of mine for one of the design of experiments classes I took in the mid to late 80's. He was very good at explaining the practical aspects of the methods as I recall.

As you go through this material, don't let yourself get bogged down worrying about the statistics. Use the formulas without thinking about it too much. After awhile, and after looking at many case studies, you will begin to see the logic behind the formulas. When you have mulitple variables, there are no easy ways to determine each of their contributions, not to mention their interactions.

You'll do fine.